Among patients treated for depression in primary care, few receive recommended intensity of treatment. Delivery of appropriate acute-phase treatment to depressed primary care patients would significantly improve clinical outcomes. Our previous research, however, suggests that inadequate treatment is more a consequence of inadequate practice organization and follow-up than deficits in provider knowledge. Effective quality improvement must assure appropriate follow-up care and improve treatment adherence. Key ingredients of a systematic practice improvement program include: "real time" monitoring of care processes, systematic assessment of outcomes, continuous feedback to providers, and support for outreach to patients who discontinue treatment prematurely. This proposal describes a randomized evaluation of such a population-based quality improvement program for depressed primary care patients. We will enroll approximately 45 primary care practices at Group Health Cooperative (a staff-model HMO) and randomize practices to 3 groups. In the "usual care" group, physicians and patients will receive no specific intervention services. In the "process feedback only group", physicians will receive reports on the progress of patients currently receiving acute-phase antidepressant treatment (medication adherence, follow-up visit frequency, specialty mental health treatment, recommendations for changes in treatment). In the "process/outcome feedback plus practice support" physicians will receive the above reports as well as data on clinical response to treatment. Physicians in this group will also be supported by a "population manager" who will assist in contacting patients to assure appropriate follow-up. A sample of patients treated by participating physicians (approximately 250 per group) will be contacted for independent, blinded assessment of clinical and functional outcomes 3 and 6 months after initiating treatment. HMO data systems will be used to assess medication use, follow-up visit frequency, and costs of care. Each of the intervention groups will be compared to the usual care group in terms of quality of care, clinical outcomes, functional outcomes, and treatment costs.